Understanding Tangent on the Unit Circle
The unit circle gives you sine and cosine values for any angle, but tangent is where things get a bit messier. You probably know the definition already: tan(theta) = sin(theta) / cos(theta). That's all it is. The tricky part is when cos(theta) equals zero, because then you're dividing by nothing and the result is undefined. Those are the angles at 90 degrees, 270 degrees, and anything coterminal with them. I learned this quickly when I was grading precalculus exams and noticed about a third of the students would write "undefined" without any reasoning, while the rest would write "infinity," which is wrong. Infinity isn't a number. Undefined is the correct answer. The graph doesn't approach infinity either — it shoots up on one side and down on the other, so calling it undefined is mathematically precise. Here's a practical way to find tangent values without crunching every angle. If you know the sine and cosine values on the unit circle, you just divide them. Let me walk through a couple of examples.
At 30 degrees, sin is 1/2 and cos is sqrt(3)/2. Divide those and the 2s cancel, leaving 1/sqrt(3). Rationalize it and you get sqrt(3)/3. That's tan(30°). At 45 degrees, both sin and cos equal sqrt(2)/2. They cancel out completely, so tan(45°) = 1. This is one of those clean results that shows up constantly in problems, so memorize it. At 60 degrees, sin is sqrt(3)/2 and cos is 1/2. The division flips those fractions, giving you sqrt(3). That's tan(60°).
The pattern at these special angles is straightforward. At 0 degrees, tangent is 0 because sine is 0. At 90 degrees, cosine is 0, so tangent is undefined. At 180 degrees, tangent is 0 again. At 270 degrees, undefined once more. At 360 degrees, back to 0. I ran into a specific problem a few years ago while working through a calculus problem set that involved finding limits of tangent near its asymptotes. The issue came up when I needed to evaluate the limit of tan(x) as x approaches pi/2 from the left side. Some textbooks treat this loosely and just say it goes to infinity, but for a rigorous analysis you need to show the left-hand limit is positive infinity and the right-hand limit is negative infinity, which means the two-sided limit doesn't exist at all. I used a small-angle substitution, letting h = pi/2 - x and then expanding tan(pi/2 - h) as cot(h), which made the behavior near zero much clearer. That trick — rewriting tangent in terms of cotangent near its asymptotes — is something I've used repeatedly in both teaching and problem-solving. It cuts through a lot of the confusion around those undefined points. One thing beginners consistently miss is that tangent has a period of pi, not 2pi. This matters because it means the tangent values repeat twice as often as sine and cosine. When you're solving an equation like tan(x) = 1, the general solution is x = pi/4 + n*pi, where n is any integer. Writing just x = pi/4 is incomplete and will cost you points on exams. The unit circle with tangent repeats every half rotation, so you need to account for all the coterminal solutions.
Get the Full Details
Another nuance that trips people up involves the sign of tangent in different quadrants. Since tangent is sine divided by cosine, it's positive whenever sine and cosine share the same sign. That happens in quadrants 1 and 3, where both are positive or both are negative. It's negative in quadrants 2 and 4. The mnemonic ASTC covers this, but it's worth understanding why rather than just memorizing it. In quadrant 3, for example, both sine and cosine are negative, and a negative divided by a negative is positive. That's why tan(210°) is positive even though the reference angle is the same as tan(30°).
When the Unit Circle With Tan Falls Short
The unit circle approach works well for exact values at standard angles, but it breaks down pretty quickly for arbitrary angles. If you need tan(37°) or tan(100°), the unit circle alone won't give you a clean answer. You'd need a calculator, which introduces rounding errors, or you'd need to use numerical approximation methods like Taylor series expansions, which converge slowly for tangent near its asymptotes. There's also the issue of vertical asymptotes. Near 90 degrees and 270 degrees, the tangent function becomes extremely sensitive to small changes in the angle. A difference of just 0.01 degrees can swing the value from positive 5700 to negative 5700. This makes tangent poorly conditioned for numerical computations, and any tool or method relying on it needs to handle that instability carefully. I've seen code crash because someone didn't account for a value landing exactly on an asymptote. If you're doing heavy trigonometric computation, especially in engineering or physics contexts, you might be better off using the sine and cosine components separately and computing tangent only when necessary. This avoids the instability and division-by-zero issues entirely until the final step. Many numerical libraries do exactly this internally.
For most students, the main takeaway is that tangent on the unit circle is simple in theory and mechanical in practice, but the edge cases are where mistakes happen. Know your special angles, remember the period is pi, watch for the undefined points, and understand why the signs change by quadrant rather than just memorizing ASTC. That last point alone will save you from a lot of avoidable errors on tests and in applied work.
